The Agent Skills Directory

[SAFE]: The skill is entirely documentation-based, providing instructional patterns for multi-agent coordination. It does not include any executable scripts, shell commands, or binary files.
[SAFE]: The allowed-tools configuration in the frontmatter restricts the agent to a minimal set of read-only tools (Read, Glob, Grep), which is a secure implementation of the principle of least privilege.
[INDIRECT_PROMPT_INJECTION]: The skill describes an orchestration pattern where a 'synthesis' step processes findings from other agents. While this constitutes a data flow between untrusted agent outputs, it is the primary intended purpose of the orchestration pattern and uses standard Markdown formatting for boundary delimitation.

parallel-agents